In computer science, database archiving refers to procedures to ensure the long-term availability of the required data relating to database systems. On the one hand, this includes simple backup/recovery procedures provided by most database management systems. On the other hand, this also includes application-oriented database archiving. Application-oriented database archiving has the following characteristics:
- Archiving of logical data granules: Based on the respective data model, the data sections to be archived can be selected individually.
- User Initiation: An archiving operation can be event-driven or explicitly initiated by a user.
- Data offloading: The operational data is physically separated from the archived data.
- Archive access: Access to the archived data must be explicitly requested if necessary. Then the data is returned to the original database.
Important aspects of database archiving are also the protection of data integrity (especially important with regard to legal requirements), the preservation of information on the origin of data and data authors (authenticity) as well as the preservation of the readability and comprehensibility of the archived data. The latter aspect is particularly critical in the archiving of databases, since the primary data itself is atomized, i.e. separated into the smallest possible, non-redundant units, and is available in different tables. In relational databases, these tables are connected to each other by key fields. If these relations between tables are lost, this leads to a loss of information, because for example, for coded fields, the connection to the table with the codes and their meaning is lost. As a result, archiving databases also differs significantly from archiving other digital objects.
Distinction Between Backup and Archiving
The term database archiving does not include backup/recovery procedures for all authors. Some authors also distinguish between the creation of backups, archiving for data offloading, and long-term archiving of databases. In the case of long-term archiving, the objective is completely different from that of backup. When backing up a database, the focus is on the short-term and fast availability of up-to-date data in the event of a malfunction or data loss from a currently operable system. Archiving for data retrieval is intended to retrieve inactive data from a currently operable database system in order to speed up queries and backup processes. Both require an actively operated database and in both cases only a copy of the data is created (and the original data is deleted in the case of offloading). Both the creation and re-import of the data are closely linked to the respective database management software. Long-term database archiving, on the other hand, aims at long-term storage of structured data that guarantees data integrity and is manufacturer-independent. For a manufacturer-independent form of storage, it is necessary to separate the data from proprietary or software-specific data formats (normalization).
---
